http://technocrat.net/d/2006/4/5/2073
Welcome to the State of Open Source press conference.

First, I'd like to make sure you all understand one thing: I dressed this
way long before I heard of Peter Quinn [Perens is wearing a suit and tie.
See http://news.zdnet.com/2100-3513_22-6054741.html]

Please enjoy lunch, which is courtesy of Prentice Hall PTR, a company that
does great things for Open Source as publisher of the "Bruce Perens' Open
Source Series": 21 books now, with all text under the Open Publication
License, with downloadable source and unencrypted PDFs. As always, the opinions
expressed here are my own, not those of Prentice or any other company.

I think my book series may be the largest producer of documentation for
Open Source software as Open Source text - by page count or book count.
Now, we're looking for authors for our next 21 books. The process, and
payment, are exactly like that of proprietary books. The revenues are the
same, because of the dynamics of book publishing. OK, that's not huge:
if you want big bucks, write bodice-rippers, not software documentation.
But most authors write for other reasons than the book income, and an
Open Source book is better for those authors than having more restrictive
rights that the publisher owns, not you. We need authors for 21 more
books, and they should contact if you'd like to write.

Please, break in with questions whenever you want.

This show is a a disappointment. It seems that Boston just isn't working
out as a venue for the Winter LinuxWorld Expo, as the Summer show in
San Francisco is huge compared to this one and of course Open Source is
bigger than ever. Most shocking about the show is the fact that the two
anchor tenants, HP and IBM, are missing from the exhibition floor. It's
New York or Bust, folks.

Let's get back to Peter Quinn. Why was there so much action around the
Massachussets Open Document format? Let's face it, folks: open file
formats are important, but they don't make it to the front page of the
Boston Globe without another agenda pushing the story. And it's easy to
underestimate how big that agenda is and how far it goes.

It's interesting to note that Jack Abramoff, the lobbyist implicated in
scandal with Republican Tom Delay, was employed by Bill Gates' dad's
law firm "Preston Gates", a political proxy for Microsoft. Microsoft
succeeded in lobbying both Republicans and Democrats to oppose ODF.

Two candidates for the Massachussets Secretary of State are already
facing off on Open Document: John Bonifaz is for it,
William Francis Galvin was one of Quinn's tormentors and remains opposed.
Departing Republican Massachussets governor Mitt Romney wants to be the
next President of the United States, and after an abortive flip-flop
on the topic, seems to be resisting pressure to abandon ODF as a means
of distancing his campaign from Microsoft's aggressive lobbying and the
Abramoff scandal.

There's a chilling effect that stems from the harassment of Quinn: other
government CIOs are being scared away from the Open Format issue because
now they know that Microsoft will do its best to end their careers if
they even try. They can see from the Abramoff scandal that Microsoft's
influence reaches the very highest level of American politics - and that
while Abramoff and Delay were damaged, Gates hasn't been touched.

GPL 3 is in the news. The original text was not usable, but that's what
you can expect from original text. I don't believe we'll see a final
license until late this year.

Linus Torvalds has publicly spoken of problem with the DRM provisions, in
my opinion jumping the gun, since there were ample opportunities for the
text to be handled on the four existing discussion committees. But Linus
doesn't like politics and policy, and thus he sort of tried to short-circuit
the process through the press without first discussing the problem with any
of the committees. That wasn't productive. Since Linus doesn't like to work
on this, I wish that instead of trying to throw thunderbolts he'd let someone
else on his team who has the right aptitude handle it.

FSF's goals regarding DRM are ambitious, and I'm not sure what the final
version will look like. They really are trying to force the issue. They
think that they can get Linus to accept a DRM text, maybe one different
from today's text. The question is whether embedded Linux is important
enough to manufacturers for them to accept a weakening of DRM in exchange
for using it. If it's not, and FSF sticks with their text and gets Linus
to back them up, they could end up falling on their swords - the DRM text
won't have a useful effect unless manufacturers who incorporate DRM actually
use Linux.

I suggested an alternate approach to DRM when we had the first GPL 3 meeting
in New York 4 years ago. My approach was that GPL 3 should require that the
user be given the information necessary to replace the software in situ - this
is of course possible because embedded devices use FLASH rather than permanent
ROM. It's actually too risky for a manufacturer to build a device with
un-modifiable software. If the user puts an un-signed kernel on a device, under
my approach the DRM would stop working, but everything else in the device
would continue to function. This strategy works today under GPL 2. But FSF
thinks they can get more done to fight pervasive DRM than I thought 4 years
ago. Let's see if they can.

We had Steve Ballmer rattle the software patent sabre against Linux in
Forbes magazine
[http://www.forbes.com/home/enterprisetech/2006/03/22/ballmer-microsoft-linux-cz_df_0322microsoft.html].
Every day that Microsoft doesn't file those suits, we get a better
defense, because of the Doctrine of Laches. That says that if you hold
off a patent suit waiting for the market to grow, you lose. Laches is an
expensive defense to run and it's no sure thing. The standard interval
for a Laches defense is 6 years, however Laches cases have been won with
a period of as little as one year.

However, it's still the case that Microsoft could stop all new innovation
in the Linux space by bringing suits using newer patents, at the cost
of some survivable damage to themselves. My theory is that they won't
do so until they get unified software patenting laws in Europe. They
won't create bad news that would sink their own legislative efforts.

The NTP vs. RIM systems case has given us a new poster-boy for
software patent problems. First, the U.S. Patent Office declared
the NTP patents invalid. Then RIM paid NTP half a Billion dollars for an
instant end to the case, because Treo sales were gaining on Blackberry
due to the doubt the case created. It's clear that justice was not
done. A classical patent troll ran a legal extortion campaign, using
patents confirmed not to embody any invention, and made a cool half Billion
from it. This
case sends a signal to get-rich-quick artists everywhere: it's not even
necessary for your patent to be valid. Become a troll, extort, litigate,
be rewarded! When you're done, take 10% of your plunder and become a
philantrophist for the tax shelter: the man on the street will consider
you a hero and a leader.

Behind NTP vs. RIM, we have the Eolas case, which is in the news again
because Microsoft has instructed its own third-party developers to stop
using the patent. Now, Eolas is a sort of self-effacing alias, it's really
the University of California and Michael Doyle, a professor at their
San Francisco campus. Doyle thinks he invented hypermedia plug-ins in 1994
[http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO1&Sect2=HITOFF&d=PALL&p=1&u=/netahtml/srchnum.htm&r=1&f=G&l=50&s1=5,838,906.WKU.&OS=PN/5,838,906&RS=PN/5,838,906],
even though they were discussed in the proposal for the HTML EMBED tag,
on an open mailing list for development of the HTML protocol more than a
year before Doyle filed his patent.
[http://www.w3.org/2003/10/301-filing.html].

The University of California is a state-owned institution containing or
operating several federal labs and is the recipient of tremendous federal
funding to employ professors and perform research. Sometimes they use that
public funding to file patents like Doyle's, and then, through companies
that they start to capitalize on their patents, they sue people. Now,
it seems especially unjust if you happen to be the defendant of a lawsuit
and have financed the plaintiff with your own taxes. Perhaps this isn't
as true when the defendant is Microsoft, a company known for its creative
use of tax loopholes. But it's true for other victims.

So, we should consider: if you pay for research with your taxes, should
you really be at peril of being sued as an infringer of the very research
you paid for? Don't you have the right to make use of that research?
Universities say no, they want to use that money to further fund the
University, and they got the Bayh-Dole Act to allow them to do that.
So, what we are actually getting in the case of the Bayh-Dole act is an
unrepresented tax on technology intended to fund Universities, and as a
probably-intended side-effect this is a full-employment act for patent
lawyers and an encouragement to litigation. Justice would be better done,
and we'd spend less, if we funded the Universities with a represented
tax and let the lawyers take care of themselves.

We also have the wildly inappropriate situation of the University of
California and its professor acting to impede the technology of the web,
essentially declaring war on an open standard: HTML.

So, after years of protest from the developer community, we saw this
year a day-long public conference including OSDL and Patent Office
officials with the goal of improving patent quality. These efforts
were encouraged and sponsored by IBM, the very company that brought the
case that created the software patent problem. In my last State of Open
Source message, I called this effort "Spitting in the wind" because I
felt that it would not be effective, and that indeed OSDL's controlling
board is made up of companies that are more interested in protecting
software patenting from Open Source than protecting Open Source from
software patents. That's still true - we haven't seen any evidence that
their efforts to improve patent quality would substantially protect
Open Source. But I was allowed to participate in a public meeting of
the effort, I was even encouraged to speak by OSDL's attorney. I brought
up some points that I'll relate to you, regarding Perjury, Triviality,
the differential capability of Litigation.

Perjury:

When one applies for a patent, an oath is
sworn. The penalty for falsehoods under that oath
[http://www4.law.cornell.edu/uscode/html/uscode18/usc_sec_18_00001001----000-.html]
in theory - is currently 5 years imprisonment, or a fine, or both. That's
in theory. Now, let's talk about practice. At the meeting, Robert Clark,
Deputy Director of the Office of Patent Legal Administration, said that
there _was_ a patent perjury case: in 1974, and that one comes along
every 25 years or so. Yet, we are aware of, for example, a much more
recent patent in which the claims included verbatim text of a published
paper by a researcher not connected with the applicant. And there are
examples like Microsoft's two-click patent, in which there does not seem
to have been any excuse for the filer to have been unaware of prior art.
There's a strong financial incentive for the unscrupulous to eavesdrop
on the open discussion lists of standards organizations or Open Source
projects and to make pre-dated patent filings with that information,
another form of perjury.

My premise in bringing this up is that there does not seem to be
any sense of peril for those who game the system. The worst that can
happen to a perjurer is that his patent claim is denied, and he can
get a continuation from the patent office and eventually get the patent
anyway. Contrast this to how courts treat perjury in non-patent matters:
impeachment of a witness for unreliable testimony is often followed by
prosecution for perjury and a lengthy jail sentence.

This creates a quandry for the patent examiner, because the total lack
of enforcement against perjury means that the examiner should not assume
that any patent application has been made in good faith. That examiner
might not get a whole day in which to evaluate a patent application,
and the fact that we can't rely on the applicant telling the truth makes
a job with an impossible time constraint even more preposterous.

Perjury is not a victimless crime: it creates intellectual _poverty_
because its victims will be unjustly denied use of technology covered
by a patent claim that, in general, they can't afford to litigate.

In some cases, the perjurer is hiding behind an attorney or a patent
agent who believes in the honesty of the claim. But the applicant should
be counseled on the peril of perjury in making an application, and the
peril should be real.

Today's policy seems to be denial that a problem exists. I submit that
improving software patent quality should improve the active pursuit
of perjurers: referral of applications from an examiner to a criminal
investigator during examination or re-examination, and we must carry
that process through to conviction on a regular basis.

This is something that the U.S. Patent Office can do without changing a
single law. The law-enforcement authorities exist, the public sympathy is
there for enforcement against white-collar crime, nobody's whining that
Ken Lay got in trouble. So, United States Patent Office, let's hear an
annoucement that you are going to be vigilant about referring perjurers
for criminal enforcement, more than four times a century.

Triviality:

Triviality is a problem for software patenting. How do we test it? Today,
the only concrete test for triviality seems to be that the claim has
not been made previously. This, obviously, promotes the claiming of
obvious art. One possible improvement in testing triviality, would be
a pre-publication jury. Functioning like a grand jury, this would be a
body of experts who are given a problem that is addressed by the claims,
without disclosure of the claim or leading questions. Can they come up
with one or more of the claims in a limited time?

Some would object to this process, proposing that the essence of novelty
in their claims comes from their simply considering a question that
nobody's thought of before. But I submit that an answer's being obvious
from the question should itself be considered as a sign of triviality.

The Penalty for Looking

While we don't create peril for the perjurers, there is peril for the
good guys - the penalty for looking - wilful infringement. While one
can defend oneself from claims of wilful infringement, the lines aren't
bright enough for engineers to apply in their day-to-day operations
with confidence, and thus our counsel gives us a simple guideline: don't
look. We need to create bright lines that engineers can follow if they
are to participate in improving patent quality.

Litigation and Cost of Litigation:

One of the largest problems faced by Open Source is the disparity in the
ability to contest a patent claim through litigation. The Open Source
participants are, in many cases, individuals who can not sustain a single
day in court but who should not be denied the right to practice technical
development. According to the American IP Association's Economic Survey,
it costs three to five Million to prosecute or defend a single case. And
thus, because only the wealthy can afford to litigate, there is only
justice for the rich. The effect of that is that patent holders can
stop an Open Source project in its tracks just by filing to sue. The
only option I see to get the Open Source developer off the hook today
is post-grant opposition through the patent office, which is is heard
or not arbitrarily depending on the pleasure of the director is, so
far, very limited in its process, and doesn't seem to have deferred
prosecution of high profile cases like RIM Systems. Only by moving the
process of contesting of a patent out of the courts does it appear that
we can grant justice to the poor as well as the wealthy.